Chuck,
AFAIK, the reference to "Nyquist frequency" in this context is about how
much data signaling speed you can get to pass through a band-limited
channel.
Not from sampling theory, but from communication theory's point of view.
Actually the original Nyquist work on this was related to telegraph
systems (1928); not from the perspective of sampling/reconstruction.
The fundamental frequency of a 'checkerboard' pattern (square waveform) is
the Nyquist frequency (= half the data symbol rate) in this context.
Regards,
Sebastian.
Sebastian E. Garcia
Slabs | Electronics development | Buenos Aires | Argentina
On Fri, December 20, 2019 2:28 pm, Chuck Corley wrote:
FYI: For NRZ the fundamental frequency is 1/2 of the Gbps bit rate.
This can be visualized as NRZ being sampled two times per fundamental
cycle, one each on what would be the 'rising' and 'falling' edges of the
fundamental frequency sine wave.
Chuck
On 2019-12-20 08:43, Chuck Corley wrote:
Good discussion below, and most of the things being said are correct.
However Nyquist continuous-time sampling theory and the frequency for
calculating copper roughness losses are two different things and
probably shouldn't be mixed.
I've worked with the very sharp people who design and build Spectrum &
Network Analyzers at HP & Agilent and National Instruments and I've done
DSP for all of those instruments. The "Nyquist rate" and "Nyquist
frequency" definitions are often confused (and also mis-applied):
"Nyquist sampling rate" = 2x frequency of interest:
"Nyquist rate" is the minimum rate you have to sample an analog signal
with an A/D converter to reconstruct a particular frequency in it
without it being aliased. The Nyquist sampling rate has to be at least
2x larger than the frequency of interest.
"Nyquist frequency" = 1/2 sampling rate:
"Nyquist frequency" is the flip side of "Nyquist rate", it's the highest
frequency in a A/D converter sampled analog signal that can be
reconstructed without aliasing. The Nyquist frequency is half of the
sampling rate.
Both of these definitions refer to Harry Nyquist's observations for
digitally sampling a continuous (analog) signal when digitally
reconstructing it using an A/D converter. Sampling an analog signal to
measure and reconstruct it with A/D measurements is different than using
NRZ or PAM-x, which are digitally sampling the waveform 1 or 0 digital
voltage levels at particular points in time against a threshold voltage
level.
For calculating dielectric or copper roughness losses the fundamental
frequency of the digital signal should be the primary reference for loss
calculations. That's the highest frequency being pushed through the
board, so that's the frequency of highest loss.
When I'm doing or talking or teaching about high speed board design, for
digital NRZ/PAM signals I usually just refer to the fundamental
frequency as my primary frequency/simulation reference.
Chuck Corley
National Instruments
On 2019-12-19 15:23, lsimonovich@xxxxxxxxxxxxxxxxxxxxx wrote:
I probably should have quantified my earlier post. Nyquist frequency is
1/2
BAUD rate. For NRZ (or PAM-2) BAUD rate is same as bit rate. It
represents
an alternating 101010.... bit pattern. So for example, at 28 GB/s
Nyquist
frequency is 14 GHz. For PAM-4 since there are 2 symbols per bit, a 56
GB/s
has a BAUD rate of 28 Gbaud, and Nyquist frequency of 14 GHz.
-----Original Message-----
From: Lee Ritchey <leeritchey@xxxxxxxxxxxxx>
Sent: 19-Dec-19 5:52 PM
To: lsimonovich@xxxxxxxxxxxxxxxxxxxxx; maruthi934@xxxxxxxxx;
shlepnev@xxxxxxxxxxxxx
Cc: si-list@xxxxxxxxxxxxx
Subject: RE: [SI-LIST] Re: Surface roughness of copper
Why is Nyquist used when referring to the clock frequency?
Lee Ritchey
Speeding Edge
P.O. Box 817
Bodega Bay, CA
94923
408-781-0253
I took the energy needed to be mad
And wrote some blues
Count Basie
Worry is like a rocking chair,
It keeps you busy, but doesn't get you anywhere.
Ross Ritchey
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx <si-list-bounce@xxxxxxxxxxxxx> On
Behalf
Of lsimonovich@xxxxxxxxxxxxxxxxxxxxx
Sent: Thursday, December 19, 2019 12:38 PM
To: leeritchey@xxxxxxxxxxxxx; maruthi934@xxxxxxxxx;
shlepnev@xxxxxxxxxxxxx
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Surface roughness of copper
When you look up roughness parameters for RTF/VLP they are very close
anyway. In fact some laminate suppliers do not offer VLP in lieu of RTF.
There is however a significant Rz difference between RTF/VLP and HVLP.
Conservatively about 1/2 depending on who's RTF foil is chosen by your
laminate supplier.
It is true modern serdes with Tx FIR and Rx DFE are capable of
tolerating
the loss. But sometimes there also requirements to meet IEEE Chip-module
and
OIF CEI VSR insertion loss budgets which have max channel loss of 10 dB
at
Nyquist. In these channels there is an expectation of only simple CTLE
equalization for the plug-in modules.
It's not unusual to have 8-10 inches of PCB routing from a switch chip
to a
faceplate SFP module for example, so even a 0.3 dB/in difference at
Nyquist,
meeting the loss budget is often a challenge.
Bert Simonovich
Signal/Power Integrity Practitioner | Backplane Specialist | Founder
Email:Lsimonovich @lamsimenterprises.com LAMSIM Enterprises Inc.
Web Site: http://lamsimenterprises.com
Blog: http://blog.lamsimenterprises.com/
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx <si-list-bounce@xxxxxxxxxxxxx> On
Behalf
Of Lee Ritchey
Sent: 19-Dec-19 2:10 PM
To: maruthi934@xxxxxxxxx; shlepnev@xxxxxxxxxxxxx
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Surface roughness of copper
Here is some measured data that you can use to decide whether smooth
copper
is worth the cost.
In 2013 we build a test PCB with a set of trace layers built from
Reverse
treat copper ( what you get if you don't specify a finish) and VLP
copper.
The paper was 5-TP5 Titled High Speed Losses in various materials
The test traces were 8" long (20 cm) and 4.5 mils (.11 mm) wide.
At 5 GHz (10 Gb/S) the loss was 0.2 db.
At 14 GHz (28 Gb/S) the loss was 2 db.
When you need to specify smooth copper depends on how good your serdes
are,
how long the path is and how good the laminate is. Some newer serdes
toleratet38 db loss at 28 Gb/S. Takes a pretty long path to need smooth
copper with such serces.
Lee Ritchey
Speeding Edge
P.O. Box 817
Bodega Bay, CA
94923
408-781-0253
I took the energy needed to be mad
And wrote some blues
Count Basie
Worry is like a rocking chair,
It keeps you busy, but doesn't get you anywhere.
Ross Ritchey
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx <si-list-bounce@xxxxxxxxxxxxx> On
Behalf
Of MARUTHI PRASANNA C
Sent: Thursday, December 19, 2019 8:18 AM
To: shlepnev@xxxxxxxxxxxxx
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Surface roughness of copper
Thank you Yuryl , I'll go through them .
On Wed, Dec 18, 2019, 8:45 PM Yuriy Shlepnev <shlepnev@xxxxxxxxxxxxx>
wrote:
Hi Maruthi,
The best way to figure it out is with a numerical experiment.
To start with, pick up a realistic roughness model that correlates
with the measurements - a number of such models for different type of
copper treatments is available in our app notes at
https://www.simberian.com/AppNotes.php Build a model of your link and
see the impact by turning the roughness on and off.
The impact of different signal degradation factors, including
roughness, for different data rates was a subject of the recent SIJ
webinar - see #9 at https://www.simberian.com/Webinars.php
Best regards,
Yuriy
Yuriy Shlepnev, Ph.D.
President, Simberian Inc.
www.simberian.com [1] [1 [1]]
Simbeor - Accurate, Productive and Cost-Effective Electromagnetic
Signal Integrity Software to Design Predictable Interconnects!
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]
On
Behalf Of MARUTHI PRASANNA C
Sent: Wednesday, December 18, 2019 2:10 AM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Surface roughness of copper
Hi everyone
What's the minimum bitrate after which I should start worrying about
the surface roughness parameters of the copper used in PCB ?
Does the amount of signal energy in the first harmonics , second
harmonics have any impact on this ?
I appreciate answers on this , thanks in advance .
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Links:
------
[1] http://www.simberian.com
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Links:
------
[1] http://www.simberian.com
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
